Position of a taper

[Sem_D220]

Hi all, it's been a while, hope you all are doing well.

thread1103-297880 is where I learned about the problematicity of applying Position to tapered features. Dean and pmarc noted that where a position tolerance is applied to a size dimension specified on the end face of a cone, as it is done in some industries, no standard provides means for a solid interpretation of whether the whole feature is required to be within the tolerance zone or just the end face (or any other local section cut where the size is directly controlled).

Can the following be a sufficient method to dismiss the ambiguity? Profile of a Surface without datum references could control the size (and form) of the cone, and a position tolerance FCF could be placed beneath the Profile FCF, similary to fig. 8-24. This way, everyone who like bonus tolerances can be satisfied (whether or not the same "bonus" for manufacturing can be achieved by other means and if it's truly beneficial here over composite profile is s topic for a completely different (old) discussion).

Everyone's input will be highly appreciated.

 

[pylfrm]

Profile of a Surface without datum references could control the size (and form) of the cone

There is no such thing as size for a conical surface in isolation. A cone is fully defined by included angle alone. Size is only meaningful for a cone in relation to something else. The profile tolerance has no datum feature references, so that is not the case here.

With the profile tolerance only really controlling form, the MMC position tolerance seems rather questionable.


pylfrm

 

[Sem_D220]

Hi pylfrm. Thanks for the reply.
I see what makes you say that the included angle fully defines a cone, but still: if a conical protrusion on a part is specified, for example - with a diameter at its base, a height dimension, and an included angle - doesn't it mean that a size requirement exists for that feature? If the datum-less Profile of a Surface tolerance allows a 0.2 equally disposed tolerance zone and every local diameter along the cone is produced 0.5 up to 0.6 mm smaller than the true profile - isn't it the size requirement that is not being met? I would even go as far as saying that in that case, the form of the feature meets the requirement but the part should be rejected because of the size of the conical feature.

Edit: just to clarify regarding the height dimension - what I had in mind is a "truncated cone" shaped feature.

 

[pylfrm]

if a conical protrusion on a part is specified, for example - with a diameter at its base, a height dimension, and an included angle

I assume you mean basic dimensions for these three parameters. If not, please clarify.

doesn't it mean that a size requirement exists for that feature? If the datum-less Profile of a Surface tolerance allows a 0.2 equally disposed tolerance zone and every local diameter along the cone is produced 0.5 up to 0.6 mm smaller than the true profile - isn't it the size requirement that is not being met?

There is no size requirement. Because the profile tolerance has no datum feature references, the true profile of the cone can be shifted axially such that the surface falls within the tolerance zone.


pylfrm

 

[Sem_D220]

Hi pylfrm,
I did mean basic dimensions, you got it right.
You make a good point about the possibility to shift the true profile and the tolerance zone axially to make any cone be in spec. as long as its included angle and form allow it. I neglected the fact that the end face is a separate feature and therefore it can't restrict the manipulation of the tolerance zone location as you described.

However, please consider the following scenario:
On the drawing view depicting the true profile of the cone, I could designate the two opposing points where the feature connects to the surface it protrudes from at the base of the cone as A and B. As described earlier, the cone would be defined with a base diameter, height dimension to the end face, and an included angle, all basic. The profile control would be specified on the conical feature "from A to B" then, the cone and the end face should be interpreted as one Complex Feature, not outside of the definition of an Irregular Feature of Size Type B (disagree?). The end face of the feature, being an integral part of it, would restrict constrain the shift of the true profile axially back and forth only to the distance allowed by the tolerance zone. (sorry, I got mixed up). This also means that for the true profile, there would be a fully defined range of diameters that vary from the base diameter to the end face diameter (the end face diameter derived from the combination of the aforementioned dimensions), this range of diameters for me is part of the size specification for the feature, and it could work with a combination of a Position tolerance relative to a datum, placed under the Profile FCF. What is your opinion?

 

[pylfrm]

Sem_D220,

I assume you have in mind something like the following, except all the dimensions are basic:
(link in case embedding fails for SVG)

The profile control would be specified on the conical feature "from A to B" then, the cone and the end face should be interpreted as one Complex Feature, not outside of the definition of an Irregular Feature of Size Type B (disagree?).

I would not consider the feature consisting of the conical surface and the flat surface at the small end to be a feature of size, because I would not say that it "may contain or be contained by an actual mating envelope".


pylfrm

 

[Sem_D220]

pylfrm,
Yes, this geometry is what I had in mind. Thank you for providing that drawing.
Would you mind to go a little bit into detail on why the conical feature including the flat end can't be contained by an actual mating envelope? I'm not insisting that it can, but - I can imagine it being contained by a conical counterpart envelope contacting it on the high points. It might be a little more difficult to gage/inspect compared to a cylindrical boss, but that's just a technical issue.

 

[pylfrm]

Sem_D220,

I wouldn't call it contained unless it is unable to escape.

Would you say that datum feature B is a feature of size in ASME Y14.5-2009 Fig. 4-29? What about in Fig. 4-30?


pylfrm

 

[Sem_D220]

pylfrm,

Let me know if I follow you regarding the "ability to escape" condition: for a cone, you can move the datum simulator axially even the smallest distance, and it will no longer be adjacent to the datum feature. For a cylinder, when you move the datum simulator axially, it still keeps contact with the datum feature, therefore, the datum feature is "unable to escape". Is that what you mean?

Datum B at fig. 4-30 is certainly not a feature of size.
Datum B at 4-29, I will avoid changing the subject, and simply conform to the "180 degrees rule": it is not a feature of size either.

The conical feature doesn't have a 180 degrees problem but it has an "escaping" problem and another thing you said earlier, is that it has no size characteristic as long as you analyze it in isolation from other features of the part.
Considering all that, the obvious conclusion is that since it is not a feature of size, it also can't be used as a feature of size datum feature to establish a datum axis. Would you confirm that conclusion?

 

[pylfrm]

Sem_D220,

The "ability to escape" condition was my rough idea for how the concept of containment should be interpreted in this context, but let me back up a bit. The definitions provided by ASME Y14.5-2009 are not sufficiently robust to provide a definitive answer here. To arrive at a reasonable answer, we can attempt to patch up some of the gaps. To that end, consider the following replacement definitions:

Unrelated actual mating envelope: A theoretical envelope outside the material, uniformly offset from a feature's true profile as far as possible in the direction toward the material. The relationship between the true profile and the actual surface is otherwise unconstrained. If the material of the feature does not provide a limit to the offset, no unrelated actual mating envelope exists.

Feature of size: A feature with an unrelated actual mating envelope.​

I think these definitions generally produce the same answers for cases that are clearly defined in the standard, and provide some additional clarity for cases that aren't. Thoughts?

Using these definitions, the conical feature (with or without the flat surface at the tip) is certainly not a feature of size. If you imagine the true profile starts out roughly aligned with the actual feature and an offset envelope progresses toward the material, it never reaches a point where it can't progress further. The feature just gets pushed away from the true profile.


If a conical surface is used as a primary datum feature as shown in Fig. 4-44, the datum feature simulator is coincident with the true profile. In this respect, a cone has more in common with a plane than it does with a feature of size like a cylinder. That doesn't mean it can't establish an axis, just that it does so in a different way.


pylfrm

 

[Sem_D220]

pylfrm,

I like your definition of the UAME, except I'm not sure regarding the requirement for it to be offset from the true profile "as far as possible in the direction toward the material". This will not be the case if an external feature is produced larger in size than its' true profile. In that case, the envelope should be required to be offset from the true profile in a direction outside the material, however, the limiting/constraining element should be the external peak points of the surface and not the air around the feature - this may be confusing.

Another thing is, that I don't consider the scenario of the conical feature "sliding out" from the envelope while the envelope keeps contracting as a valid way to establish an unrelated actual mating envelope from the feature. The actual mating envelope needs to cover the entire feature from the base to the small end. When it keeps contracting to the point that it looses the area of the feature near the base, and when there is an axial relative movement between the conical feature and the theoretical/physical datum feature simulator or the simulator of the UAME, it no longer constitutes an actual mating envelope. So, per my point of view, there is a limited amount of radial contraction that any conical feature allows for a counterpart envelope or datum simulator until the actual mating envelope is established, or the datum axis is ready to be derived.

If for the conical datum feature from fig. 4-44 the datum feature simulator is coincident with the true profile as you say, then the datum axis is derived independently from the geometry of the actual surface of the datum feature. Are you sure about that?

 

[pylfrm]

I'm not sure regarding the requirement for it to be offset from the true profile "as far as possible in the direction toward the material". This will not be the case if an external feature is produced larger in size than its' true profile.

I intended "in the direction toward the material" to specify which extreme should be sought, not to specify that the envelope should be on a certain side of the true profile. I definitely need to find a better way to phrase that. Nevertheless, I think we're basically on the same page here.

The actual mating envelope needs to cover the entire feature from the base to the small end.

Although I failed to mention it, complete coverage of the feature should probably be part of the definition for the proposed envelope. For a width, you would have two infinite parallel planes. For a hole or shaft, you would have a cylindrical surface of infinite length. For a cone, you would have a conical surface extending from the apex off to infinity in whichever direction is relevant.

So, per my point of view, there is a limited amount of radial contraction that any conical feature allows for a counterpart envelope or datum simulator until the actual mating envelope is established, or the datum axis is ready to be derived.

If we're talking about a conical surface only, changing the offset between the true profile and the envelope does not change the geometry of the envelope. If the envelope geometry doesn't change, how can there be a limit to the offset?

If for the conical datum feature from fig. 4-44 the datum feature simulator is coincident with the true profile as you say, then the datum axis is derived independently from the geometry of the actual surface of the datum feature. Are you sure about that?

I'm not sure what you're getting at here. Contact between the actual surface and the datum feature simulator establishes the relationship between the part and the true profile, just like it does with a planar feature. Perhaps a better way to explain what I meant is to say that the [BSC] modifier (shown in Fig. 4-31(b)) is the implied default for conical primary datum features and for planar primary datum features.


pylfrm

 

[Sem_D220]

If we're talking about a conical surface only, changing the offset between the true profile and the envelope does not change the geometry of the envelope. If the envelope geometry doesn't change, how can there be a limit to the offset?

If I was to simulate an unrelated actual mating envelope deriving it from the cone depicted in the embedded figure you provided earlier using a physical gage, I would use an expanding/contracting device with a conical cavity and keep as little clearance as possible between the face of the simulator where the opening is and the face of the cylindrical feature of diameter A in the part. I would contract the gage around the feature radially until it can't contract anymore while making sure there is no axial relative movement between the feature and the gage, keeping the face clearance minimal all the time.
However, given the infinite theoretical datum feature simulators you mentioned - I think I understand the point you convey. But, I don't see how the infinity requirement is implied by the standard, or how this could be beneficial or practical in real life.

I'm not sure what you're getting at here. Contact between the actual surface and the datum feature simulator establishes the relationship between the part and the true profile, just like it does with a planar feature. Perhaps a better way to explain what I meant is to say that the [BSC] modifier (shown in Fig. 4-31(b)) is the implied default for conical primary datum features and for planar primary datum features.

Perhaps we don't interpret the term "true profile" exactly the same way. As I understand it, the true profile is the exact form, size, and orientation geometry of the feature as defined by basic dimensions in the drawing. As far as I understand, contact between datum feature simulator and the datum feature establishes the relationship between the part and the datum, and it is in no way related directly to the true profile. This is probably the source of misunderstanding here. Please tell me where I'm wrong. I also don't completely follow you regarding the default BSC requirement for primary conical and planar datum features. The BSC modifier in fig. 4-31 (b) is related to the location of the datum feature simulator (Edit: relative to higher precedence datums). I can see how the location is relevant for a secondary datum feature simulator like the one discussed in that figure, but I don't understand the meaning of basic location of a datum feature simulator in the context of a primary datum feature, be it a cone or a planar surface. Sorry for asking repeatedly for clarifications, but could you explain?

 

[pylfrm]

I would contract the gage around the feature radially until it can't contract anymore while making sure there is no axial relative movement between the feature and the gage, keeping the face clearance minimal all the time.

Consider a slight modification of the example image I posted: Imagine a cylindrical surface of diameter 0.8*F is added between the conical surface and the adjacent flat shoulder to create an undercut. How would you proceed in that case?

But, I don't see how the infinity requirement is implied by the standard, or how this could be beneficial or practical in real life.

The standard doesn't specifically imply infinite extent of envelopes, boundaries, or datum feature simulators, but it certainly does imply sufficient extent to cover the entire actual feature. The alternative would be to ignore some portion of the feature, and I hope you agree that's a can of worms that shouldn't be opened.

If the extent is sufficient to cover the entire actual feature, then it really isn't relevant where the 'edges' of the envelope fall beyond that. In fact, the 'edges' may as well not exist. Eliminating them leaves an infinite envelope, and one less distraction.

As I understand it, the true profile is the exact form, size, and orientation geometry of the feature as defined by basic dimensions in the drawing.

No disagreement there. For the sake of discussion, imagine a drawing that fully defines the true profile of the entire part. Datum feature references constrain the spatial relationship between the actual part and the true profile, so datum feature simulators are necessarily related to both.

Fig. 4-28 is an example of the [BSC] modifier explicitly applied to a primary datum feature reference. If you think of it as specifying the relationship between the datum feature simulator and the true profile, then the distinction between primary and lower-precedence datum feature references disappears.

Fig. 4-28 relates to para. 4.13 regarding mathematically defined surfaces, and presumably the main goal was to define the procedure for features too complicated to specify in a 2D drawing. Most of the concepts seem equally applicable to simpler features though. Planes and cones can certainly be mathematically defined.


pylfrm

 

[Sem_D220]

pylfrm,
Thank you for clarifying about how the BSC modifier can be applied to primary datums, and about your view on the relationship between the true profile and the part established by datum features.
I think that after looking from this perspective, I understand figure 4-28 better.

I am trying to accept the idea that a conical datum feature acts more like a planar datum feature or a mathematically defined surface (like the one in fig. 4-28) than like a feature-of-size datum feature, but I'm still struggling
To answer your question about the modified example with the undercut added - the modified version is certainly more favorable over the original. If the undercut allows some space for the envelope simulating device to overhang from the conical surface and cover all of it - we get a better representation of the surface - be it a datum simulation process or simulation of the UAME for other purposes. I would do the simulation exactly the same way I described earlier, except I wouldn't have to worry about leaving as little space as possible between the gage and the flat face of the feature near the cone, or maybe I would have to worry about it less. As a side note, I suppose that I don't have to tell you that if this is a machined part and the conical feature is important and accurate, an undercut like you described is going to be there anyway. Either way with the undercut or without it, although I'm far from being fluent in gaging and measurement techniques, I think that technically it is possible to contract a device around the cone radially until it is constrained physically by the surface of the conical feature, as long as we are able to prevent axial movement between the simulator and the part. If there is a difficulty there - it is only technical. This is not something that doesn't make sense geometrically. And again, since any conical feature similar to the one in your example figure has well defined and measurable range of diameters spanning from the base till the end face, decreasing linearly along the axis, I do not see how a cone is a sizeless feature that can be compared to a flat one. Another reference for this discussion can be the last sentence of the text under the "Means this" section of fig. 8-17. It says "the surface must be within the specified limits of size". The profile control for that surface, just like in my initial suggestion in this thread, doesn't call out any datums, so it can hardly be argued that the "size of the surface" specification (Edit: which is provided here by the combination of plus-minus diameter on the base and form and included angle refinement through Profile) is "in relation to something else". Fig. 8-17 describes a very simple case, and I hope it is not a controversial figure.

 

[pylfrm]

I would do the simulation exactly the same way I described earlier, except I wouldn't have to worry about leaving as little space as possible between the gage and the flat face of the feature near the cone, or maybe I would have to worry about it less.

We agree that a conical primary datum feature is a valid concept, so I'll concentrate on that aspect for the moment. Assume we take advantage of the undercut and use a datum feature simulator that extends beyond the actual datum feature at both ends. As long as there is some overhang and no interference with other features, it doesn't matter where the ends of the simulator are. We can ignore the ends and concentrate on the conical surface of the simulator, because that's all that actually interacts with the part.

Now imagine we radially contract the datum feature simulator surface. This does not actually change the geometry of the surface itself; we still have a conical surface with an included angle of 60 degrees. A radial contraction of x is exactly equivalent to an axial translation of x*sqrt(3). Do you agree?

This equivalence means that the datum feature simulator may as well be a solid piece of material. A conical socket with an opening diameter of roughly 0.9*F should do the job. Datum simulation would basically be a matter of pushing the part up against the simulator axially.

I do not see how a cone is a sizeless feature that can be compared to a flat one.

Consider the behavior of a conical feature with an included angle of 178 degrees or so. Don't you see some strong similarities between that, Fig. 4-28, and a flat feature?

Another reference for this discussion can be the last sentence of the text under the "Means this" section of fig. 8-17. It says "the surface must be within the specified limits of size". The profile control for that surface, just like in my initial suggestion in this thread, doesn't call out any datums, so it can hardly be argued that the "size of the surface" specification (Edit: which is provided here by the combination of plus-minus diameter on the base and form and included angle refinement through Profile) is "in relation to something else".

In Fig. 8-17, I think the first "Means this" sentence provides a clear and complete explanation of the meaning of the profile tolerance. I don't think the same can be said for the second sentence and the toleranced dimension though. The most similar example I see in the standard with a decent explanation is Fig. 2-21. It shows an axial toleranced dimension of 9.9-10.1 and a diametral basic dimension of 23, but I think an axial basic dimension of 10 and a diametral toleranced dimension of 22.97-23.03 would result in essentially the same meaning. From there it's not much of a leap to change the axial basic dimension to zero and the diametral toleranced dimension to 25.97-26.03, which yields essentially the same scheme as Fig. 8-17. Another option would be to make all the dimensions basic and apply a profile tolerance of 0.029668 to the conical surface, using the large flat surface as the primary datum feature reference. In all of these cases, the "something else" is the flat surface. Without that, how can the notion of size be meaningful?


pylfrm

 

[Sem_D220]

You make a good point about the possibility to use a rigid datum feature simulator to establish a datum from a conical datum feature. Although a physical datum feature simulator for a cone with given dimensions still has to be of some minimum base diameter and a maximum small-end diameter to be valid, I agree that the size parameter here is questionable, and not similar to that of cylindrical or rectangular features of size.

The conclusion is that cones are not features of size. They can be used to establish a datum axis, but can't be controlled for location and orientation by Position controls.

Can an analogical conclusion be made about flat tapes?

 

[pylfrm]

Can an analogical conclusion be made about flat tapes?

Meaning a pair of nonparallel flat surfaces, such as shown in Fig. 2-20? Sure.


pylfrm

 

[Sem_D220]

Thank you pylfrm.
I was thinking of a symmetrical flat taper (equal angles between the base and the 2 angled surfaces), I suppose it's not a feature of size either. How about a local cross-section controlled with Position? Sometimes a toleranced dimension is defined at a basic location along a tapered feature, from a datum perpendicular to the center plane of the flat taper. The toleranced dimension is associated with Position control. The tolerance zone for the controlled center in this case is between 2 parallel lines.

 

[pylfrm]

How about a local cross-section controlled with Position? Sometimes a toleranced dimension is defined at a basic location along a tapered feature, from a datum perpendicular to the center plane of the flat taper.

Based on the example of Fig. 2-19, the toleranced dimension might be considered to control the entire feature, not just one cross-section.

That said, it would be easy enough to include a note clarifying that the tolerances only apply in the specified plane. As long as that plane is fully defined in the same DRF as the position tolerance then I don't see why this scheme couldn't be viable. I'm not sure how useful it would be though. Do you have any applications in mind?


pylfrm